Point Cloud Library (PCL)  1.14.0-dev
distances.h
1 /*
2  * Software License Agreement (BSD License)
3  *
4  * Point Cloud Library (PCL) - www.pointclouds.org
5  * Copyright (c) 2010-2011, Willow Garage, Inc.
6  * Copyright (c) 2012-, Open Perception, Inc.
7  *
8  * All rights reserved.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  *
14  * * Redistributions of source code must retain the above copyright
15  * notice, this list of conditions and the following disclaimer.
16  * * Redistributions in binary form must reproduce the above
17  * copyright notice, this list of conditions and the following
18  * disclaimer in the documentation and/or other materials provided
19  * with the distribution.
20  * * Neither the name of the copyright holder(s) nor the names of its
21  * contributors may be used to endorse or promote products derived
22  * from this software without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
25  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
26  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
27  * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
28  * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
29  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
30  * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
31  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
32  * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
34  * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
35  * POSSIBILITY OF SUCH DAMAGE.
36  *
37  * $Id$
38  *
39  */
40 
41 #pragma once
42 
43 #include <Eigen/Core>
44 
45 #include <algorithm>
46 #include <cstring>
47 #include <vector>
48 
49 namespace pcl {
50 namespace distances {
51 
52 /** \brief Compute the median value from a set of doubles
53  * \param[in] fvec the set of doubles
54  * \param[in] m the number of doubles in the set
55  */
56 inline double
57 computeMedian(double* fvec, int m)
58 {
59  // Copy the values to vectors for faster sorting
60  std::vector<double> data(fvec, fvec + m);
61 
62  std::nth_element(data.begin(), data.begin() + (data.size() >> 1), data.end());
63  return (data[data.size() >> 1]);
64 }
65 
66 /** \brief Use a Huber kernel to estimate the distance between two vectors
67  * \param[in] p_src the first eigen vector
68  * \param[in] p_tgt the second eigen vector
69  * \param[in] sigma the sigma value
70  */
71 inline double
72 huber(const Eigen::Vector4f& p_src, const Eigen::Vector4f& p_tgt, double sigma)
73 {
74  Eigen::Array4f diff = (p_tgt.array() - p_src.array()).abs();
75  double norm = 0.0;
76  for (int i = 0; i < 3; ++i) {
77  if (diff[i] < sigma)
78  norm += diff[i] * diff[i];
79  else
80  norm += 2.0 * sigma * diff[i] - sigma * sigma;
81  }
82  return (norm);
83 }
84 
85 /** \brief Use a Huber kernel to estimate the distance between two vectors
86  * \param[in] diff the norm difference between two vectors
87  * \param[in] sigma the sigma value
88  */
89 inline double
90 huber(double diff, double sigma)
91 {
92  double norm = 0.0;
93  if (diff < sigma)
94  norm += diff * diff;
95  else
96  norm += 2.0 * sigma * diff - sigma * sigma;
97  return (norm);
98 }
99 
100 /** \brief Use a Gedikli kernel to estimate the distance between two vectors
101  * (for more information, see
102  * \param[in] val the norm difference between two vectors
103  * \param[in] clipping the clipping value
104  * \param[in] slope the slope. Default: 4
105  */
106 inline double
107 gedikli(double val, double clipping, double slope = 4)
108 {
109  return (1.0 / (1.0 + pow(std::abs(val) / clipping, slope)));
110 }
111 
112 /** \brief Compute the Manhattan distance between two eigen vectors.
113  * \param[in] p_src the first eigen vector
114  * \param[in] p_tgt the second eigen vector
115  */
116 inline double
117 l1(const Eigen::Vector4f& p_src, const Eigen::Vector4f& p_tgt)
118 {
119  return ((p_src.array() - p_tgt.array()).abs().sum());
120 }
121 
122 /** \brief Compute the Euclidean distance between two eigen vectors.
123  * \param[in] p_src the first eigen vector
124  * \param[in] p_tgt the second eigen vector
125  */
126 inline double
127 l2(const Eigen::Vector4f& p_src, const Eigen::Vector4f& p_tgt)
128 {
129  return ((p_src - p_tgt).norm());
130 }
131 
132 /** \brief Compute the squared Euclidean distance between two eigen vectors.
133  * \param[in] p_src the first eigen vector
134  * \param[in] p_tgt the second eigen vector
135  */
136 inline double
137 l2Sqr(const Eigen::Vector4f& p_src, const Eigen::Vector4f& p_tgt)
138 {
139  return ((p_src - p_tgt).squaredNorm());
140 }
141 } // namespace distances
142 } // namespace pcl
double l2Sqr(const Eigen::Vector4f &p_src, const Eigen::Vector4f &p_tgt)
Compute the squared Euclidean distance between two eigen vectors.
Definition: distances.h:137
double gedikli(double val, double clipping, double slope=4)
Use a Gedikli kernel to estimate the distance between two vectors (for more information,...
Definition: distances.h:107
double huber(const Eigen::Vector4f &p_src, const Eigen::Vector4f &p_tgt, double sigma)
Use a Huber kernel to estimate the distance between two vectors.
Definition: distances.h:72
double computeMedian(double *fvec, int m)
Compute the median value from a set of doubles.
Definition: distances.h:57
double l2(const Eigen::Vector4f &p_src, const Eigen::Vector4f &p_tgt)
Compute the Euclidean distance between two eigen vectors.
Definition: distances.h:127
double l1(const Eigen::Vector4f &p_src, const Eigen::Vector4f &p_tgt)
Compute the Manhattan distance between two eigen vectors.
Definition: distances.h:117